• Please review our updated Terms and Rules here

North Star Horizon power

Ttpilot

Experienced Member
Joined
Nov 8, 2020
Messages
100
Location
South Dakota USA
This is to follow up on a thread I started a while back about my dead North Star Horizon motherboard. After repairing the traces and a bad resistor, I decided to replace the original linear power supply with a switching power supply, specifically a Meanwell HRP-75-7.5 and two RS-15-15s. That's what I'm using to power my Altair 8800c. Each power supply checks out okay when I measure the voltages, which are calibrated to +8v, +16v, and -16v. When plugged into the quick connect lugs on the motherboard I get a strange result. The +8v supply is fine, including the +5v supply derived from it. However, neither 16v line shows any life, either at the lugs or anywhere else in the circuit. They're plugged into the correct quick connect lugs, and I know power is coming out of the power supply units, so I'm confused as to why nothing shows up on the motherboard

Suggestions welcome!
 
Looking at the power supply specs, the HRP-75-7.5 is rated for 7.5v, 10a, 75w. The RS-15-15 is rated for 15v, 1a, 15w. Maybe that's too weak to supply enough power to that motherboard? The RS-100-15 delivers 15v, 7a, 105w. Should I have used that one?
 
Do you have the typical power consumption at the rated voltages for the various components of your system? This information can usually be found in the various manuals.

If you sum the individual consumptions up for each voltage rail, that gives you the minimum requirement from the power supply.

The other way of doing it is to look at what each voltage regulator is rated at - and sum those together.

If the power supply is not capable of supplying the requisite current - it will probably shut down.

Dave
 
If the power supply is not capable of supplying the requisite current - it will probably shut down.

That seems to be what is happening. Interestingly, that same combination powers a 9-slot backplane plus 4 or 5 cards on my Altair 8800c. The Horizon mb must be drawing a lot more power to supply on-board logic and i/o, cards (none installed at the moment), and two 5.25” floppy drives
 
Ah, the dreaded words “floppy drives”. These are fairly power-hungry at +12V. What make and model are they?

Dave
 
They are Panasonic JU-455s. The drives aren’t attached at the moment, although I used them in another system recently
 
About a year ago, I brought a N* back to life. It had 2 12V regulators in parallel. Both were bad. The blue tantalum cap was bad as well. I just cut the input filter on one of the drives on one of the leads ( it still dangles there but at least it is not loading the power line down).
I forget which supply line it was but one of the 1/4 in slip on connectors had a bad crimp and wouldn't even allow an single ampere to be drawn through it. ( I'm not a fan of these crimp connectors ). I took my solder iron to it. The plastic cover didn't like it but I'm more interested in function over beauty.
After that the only other problem was one of the memory boards blew a really large tantalum cap that did a little fire damage to the board before I could shut it down.
I later was able to read disk and format one disk without issues.
I do like the N* machines. The general quality was quite good.
Dwight
 
Yes, Dwight is 100% correct - have you checked the motherboard voltage regulators and capacitors for signs of short circuits? This would be a very important test before powering up the motherboard (especially if it showed signs of damaged PCB tracks).

The motherboard regulators themselves would probably have shutdown if the output capacitors of the regulators had gone short circuit. However, there are capacitors on the input of the motherboard voltage regulators. If one of these had gone short circuit, that would cause your external supply to shutdown.

Dave
 
This is going to sound stupid, but I’ve got a case of information overload, I guess. How do I check these (meaning the 7812s, 78L12, 79L12, and the relevant caps) while they’re on the board with no power supplied?
 
Nothing sounds stupid here...

Let's take it step by step then...

Use a multimeter (set to measure resistance - say a full scale range of 1 kOhm) and (with the power OFF and power supplies disconnected from the motherboard) measure the resistance from GND to each of the incoming supplies on the connectors.

See what the readings are and post them. We are looking for a permanently 'low' resistance, indicating a short circuit (or leaky) capacitor or a short circuit (input to GND) voltage regulator to start with.

Does this make sense to start with?

Dave
 
Yes it does, thanks. As you can see, my diagnostic skills with circuits are not good. I’d much rather debug code lol. Anyway, I hooked up to one of the ground lugs and then measured on the power lugs:

8v to ground: 100 ohms

-16v to ground: 1.6 ohms

+16v to ground: 1.6 ohms
 
Last edited:
Oh, those 1.6 Ohm readings don’t look good...

From Ohms Law, that would give us round about 8A!

I would suggest now removing the capacitors from the input side of the voltage regulators to see if that resolves the issue with your multimeter readings. If it does, replace the capacitors.

Note that these capacitors are polarised, so note any + or - markings on the device (or PCB) before removal and ensure they go back the same way round.

If possible, only remove one lead of the capacitor. Sometimes this is easy, other times not (depending upon the capacitor lead arrangement - axial or radial).

Once you have the capacitor out of circuit (or one lead disconnected) you can test it. Ask if you need further details on this.

Dave
 
The fact that both the + and - 16V rails look shorted, may mean that someone in the past connected the wrong polarity to the boards. This would quickly take out the input filters ( puff ).
When you get to the point that the input filters are good ( tested or replaced ). Then you can look at the regulator outputs. Anyway, follow Dave's instructions. He is a well respected trouble shooter.
One side thing. What type of multi-meter do you have? If it is digital the polarity of the leads in the Ohms position is the same as for reading voltages. If it is an analog meter, most of the less expensive ones swapped the polarity in the ohms. The better quality one didn't. It is always worth checking if you have an analog meter. Always use the correct polarity when testing power leads, with an Ohm meter. 1.6 ohms sounds like you might be looking at diodes on the board with the wrong polarity.
Dwight
 
The fact that both the + and - 16V rails look shorted, may mean that someone in the past connected the wrong polarity to the boards. This would quickly take out the input filters ( puff ).
When you get to the point that the input filters are good ( tested or replaced ). Then you can look at the regulator outputs. Anyway, follow Dave's instructions. He is a well respected trouble shooter.
One side thing. What type of multi-meter do you have? If it is digital the polarity of the leads in the Ohms position is the same as for reading voltages. If it is an analog meter, most of the less expensive ones swapped the polarity in the ohms. The better quality one didn't. It is always worth checking if you have an analog meter. Always use the correct polarity when testing power leads, with an Ohm meter. 1.6 ohms sounds like you might be looking at diodes on the board with the wrong polarity.
Dwight
I have an AstroAI DM6000AR multimeter

I assembled the system in 1978 and it worked fine until the mid 1980s, when I shelved it. I got it out a couple of years ago and turned it on. It ran fine for a week or so, then quit working. The power supply was putting out the correct voltages, although the -16 rail kept blowing up when I applied the multimeter on the edge connectors. I turned to some other projects for a year or so and finally got back to trying to fix it. I recently pulled the old linear power supply and replaced it with several Meanwell switching units. With the original supply, I always felt like the neighborhood lights dimmed when I turned it on

I'm not aware that I've ever crossed the polarity when hooking the system up
Oh, those 1.6 Ohm readings don’t look good...

From Ohms Law, that would give us round about 8A!

I would suggest now removing the capacitors from the input side of the voltage regulators to see if that resolves the issue with your multimeter readings. If it does, replace the capacitors.

Note that these capacitors are polarised, so note any + or - markings on the device (or PCB) before removal and ensure they go back the same way round.

If possible, only remove one lead of the capacitor. Sometimes this is easy, other times not (depending upon the capacitor lead arrangement - axial or radial).

Once you have the capacitor out of circuit (or one lead disconnected) you can test it. Ask if you need further details on this.

Dave
Thanks! I'll do that as soon as I have a chance
 
I have been having a second look at the schematics (last page of http://cini.classiccmp.org/pdf/Northstar/ns_hrzd_manual.pdf).

I assume (?) all of the large “external to the motherboard” capacitors have also been replaced by the meanswell’s? If not, obviously check these first.

The input capacitors (C7, C3 and C1) are all tantalum bead types (as are C10, C4, and C2).

As this machine is your own (and I assume you plan on keeping it) I would be inclined to just replace these “smurf grenades”. Replace them with the same value with at least the same (or higher) voltage rating. Note that a higher voltage rating will be in a larger package - so just be careful that you have adequate space to install them.

Tantalum bead capacitor polarisation is usually identified by a ‘+’. Installing these backwards will cause a localised volcano effect!

EDIT: I see the parts list doesn’t identify the maximum working voltage for these capacitors. I would go for at least twice the corresponding voltage rating of the Meanwell DC output voltage (subject to any space constraints). They are dirt cheap anyhow if you order the wrong part...

Of course, I would still test the original capacitors (even if you plan to replace them anyhow) to see if one (or both) are shorted. You still need to identify the faulty bit(s)!

Dave
 
Last edited:
I am going to assume that you have removed all of the S100 cards from the motherboard.

The only devices I can see on the motherboard that uses the + and - 12 Volts are the 1488 and 1489 RS232 buffer ICs. I assume these are in IC sockets?

Both the + and - 12 Volt regulators are the ‘L’ type (so somewhat tiny) and this supports my theory that they are very lightly loaded. Also the 'L' variant is rated at 100 mA max, so a short circuit on the output side would have caused the motherboard 'L' regulator to have shutdown and not the Meanwell.

Just thinking ahead to the next tests...

Just out of interest, how have you connected the Meanwell’s to the N*? Any chance of doing a very simple line diagram (or even a scanned hand drawing).

Dave
 
Last edited:
I was also wondering, since you stated you're better with coding, did you reverse the leads for the minus supply? To supply negative voltage to the board you must tie the plus output lead from the Meanwell to common (AKA ground) and negative lead to the negative input of the board. You do this to make the board common (AKA ground) the positive side of the power supply feeding the minus side. Do not tie the negative lead to common (ground) in this case.
 
I was also wondering, since you stated you're better with coding, did you reverse the leads for the minus supply? To supply negative voltage to the board you must tie the plus output lead from the Meanwell to common (AKA ground) and negative lead to the negative input of the board. You do this to make the board common (AKA ground) the positive side of the power supply feeding the minus side. Do not tie the negative lead to common (ground) in this case.
Yes, they’re reversed. I’ve used the same setup for my Altair 8800c. In this case I even labelled each lead
 
Also, start by disconnecting everything and measure the voltage with the minimum connected. You want just the mother board first.
As for lights dimming, A big transformer with limited overheard on magnetic material will often saturate if the AC is at 0 volts when turned on ( crossing 0 ). For inductive loads, like transformers, it is best to turn them on at the peak of the cycle. For capacitive loads it is best to turn on at zero crossing. A lot of engineers get this messed up.
I'm with Dave on the tantalums. If you don't have the technics to find them they are so cheap you might as well replace them. I recommend using a sharpie to mark the board for the + lead before removing them if the boards are not already marked.
Also, the switchers may not like starting from a low voltage of a variac turned down. Many switchers use the output rail to power the control logic and have a pulse circuit to drive the transformer with one cycles to get things started. If you start with a reduced voltage, they may not start.
Use the variac for linear supplies only.
Dwight
 
Back
Top